Community
Participate
Working Groups
As a regression of https://bugs.eclipse.org/bugs/show_bug.cgi?id=549855 I observe several NPE issues when parsing module-info files of OpenJDK, e.g. http://hg.openjdk.java.net/jdk/jdk/file/fb39a8d1d101/src/jdk.zipfs/share/classes/module-info.java http://hg.openjdk.java.net/jdk/jdk/file/fb39a8d1d101/src/java.base/share/classes/module-info.java For instance there is: java.lang.NullPointerException at org.eclipse.jdt.internal.compiler.ast.Javadoc.resolveReference(Javadoc.java:426) at org.eclipse.jdt.internal.compiler.ast.Javadoc.resolve(Javadoc.java:293) at org.eclipse.jdt.internal.compiler.ast.CompilationUnitDeclaration.resolve(CompilationUnitDeclaration.java:645) at org.eclipse.jdt.internal.compiler.Compiler.process(Compiler.java:901) at org.eclipse.jdt.internal.compiler.ProcessTaskManager.run(ProcessTaskManager.java:145) at java.base/java.lang.Thread.run(Thread.java:830) The general problem is that a module declaration's scope is a MethodScope without an Enclosing ClassScope. So far, the assumption in org.eclipse.jdt.internal.compiler.ast.Javadoc in several places is that there is some parent ClassScope somewhere and null is not an option. I'd like to propose a fix which checks for null at some more places. It would only be a short-term band-aid. We should generally consider whether a module declaration's scope should be a MethodScope and not something else (see https://bugs.eclipse.org/bugs/show_bug.cgi?id=553472) Maybe also in the context of https://bugs.eclipse.org/bugs/show_bug.cgi?id=548069. I'd hope that my proposed fix can make it into RC2 - otherwise it would be a nasty regression for the OpenJDK usecase.
New Gerrit change created: https://git.eclipse.org/r/154079
Unfortunately RC2 has already sailed, see bug 553793. Only in cases of severe damage to life or property we could request a respin. As a workaround you may have to disable javadoc analysis in the preferences. Did you see that Vikas created bug 553564 to follow up from bug 553472 in what I called the mid-term solution?
(In reply to Stephan Herrmann from comment #2) > Unfortunately RC2 has already sailed, see bug 553793. > > Only in cases of severe damage to life or property we could request a respin. > Oh, too bad. > As a workaround you may have to disable javadoc analysis in the preferences. > Did you see that Vikas created bug 553564 to follow up from bug 553472 in > what I called the mid-term solution? Ah, thanks for that hint. Maybe I can try to have a look at this. It should probably be a predecessor to further work on 548069. In any case, please have a look and check if this regression fix can be merged until we have anything better...
The code change in gerrit looks good for now (we should revisit it after bug 553564 if all that will still be needed). Would you be able to also provide a test case? Jenkins build has been retriggered.
(In reply to Stephan Herrmann from comment #4) > The code change in gerrit looks good for now (we should revisit it after bug > 553564 if all that will still be needed). > > Would you be able to also provide a test case? > > Jenkins build has been retriggered. I've uploaded a new patch set with a test that demonstrates the NPE if run without the fixes. The second build did not show errors.
> I've uploaded a new patch set with a test that demonstrates the NPE if run > without the fixes. -> that is, one of the 3 potential NPEs...
My view is that we should do a re-spin for this issue. In the current context, it would probably means joining the re-spin.
Going forward, can we add test cases for all the javadocs in module-info files of openJDK? ( probably warrants another bug)
New Gerrit change created: https://git.eclipse.org/r/154159
(In reply to Eclipse Genie from comment #9) > New Gerrit change created: https://git.eclipse.org/r/154159 R4_14_maintenance branch.
We can put this fix in the next 4.15 I build. >>As a workaround you may have to disable javadoc analysis in the preferences. This can be used for 4.14
Gerrit change https://git.eclipse.org/r/154079 was merged to [master]. Commit: http://git.eclipse.org/c/jdt/eclipse.jdt.core.git/commit/?id=3ff9186163506e12e3283335bc32f49dad643c22
Ideally the change in Javadoc.java should be reverted once Bug 553564 is fixed. The next I build should be usable fir OpenJDK usecase
